Common Cause Abduction and the Formation of Theoretical Concepts in Science GERHARD SCHURZ ABSTRACT. Abductions are conceived as special pattcms of inference to the best explanation whose structure determines a particularly promising abduc- tive conjecture (conclusion) and thus serves as an abductive search slrategy (§J). An important distinction is that between selective abductions, which choose an optimal candidate from a given multitude of possible explanations, and creative abductions, which introduce new theoretical models or concepts (§2). The paper focuses on creative abductions, which are essential for sci- entifk progress, although they arc rarely discussed in the literature. It is sug- gested to demarcate SCientifically fruitful abductions from purely speculative abductions by the criterion of cm/sa/llnification (§3). Based on various his- torical examples it is demonstrated that C0ll11110n calise abduction from corre- lated dispositions is the fundamental abductive operation by which new the- oretical concepts are scientifically generated (§4). Statistical factor analysis can be regarded as a statistical generalization of common cause abduction (§5). When scientists start to develop theoretical models of their conjectured common (unobservable) causes, common cause abduction turns into what is called theoretical model abduction (§6). 1 On the relation between inference to the best explanation and abduction Harman [21] understood inference to the best explanation (!BE) and abduction as more or less equivalent. Both inferences serve the goal of inferring something about the unobserved causes or explanatory reasons of the observed events. This was also the understanding of abduction in the mind of the inventor of 'abduction', C.S. Peirce (cf. [47, §5.189]). Nevertheless, I suggest to make a difference here. By an 'inference' I mean a certain (logically explicable) schematic pattem which specifies the conclusion as a (syntactical) jimct/on of the premises. In the case of an abductive inference, the premises describe the phenomenon which is in need of explanation, possibly together with hackground knowledge of a certain form, and the conclusion is an abductive conjecture which is set out to further test operations. However, in the case of IBE no such pattern exists. Rather, the space of possible explanatory hypotheses and their evaluation must already be given in the premises,